Estimating and locating the volumes of multiple primary non-small cell stage-I bronchogenic carcinomas from medical images using specially devised morphological operations

摘要

Lung tumour segmentation is one of the challenging tasks in medical image processing due to the high diversity in appearance of tumour tissue among different cases. In this paper, a novel morphological based semi-automatic method for estimating and locating the volumes of multiple primary non-small cell stage-I bronchogenic carcinomas in medical CT DICOM images is proposed. Various morphological image processing techniques along with a sequence of other image processing methods are used. Experiments are carried out on real datasets from various CT DICOM images of lung tumours. The results verify that this method is promising and accurate with the sensitivity of 93.3% and the specificity of 90.0%.